The RSI allows two Replication Servers to exchange transactions across a route. The RSI consists of:
A stable queue and thread at the primary Replication Server. The thread reads the stable queue and writes to the network.
A thread at the replicate Replication Server. This thread reads from the network and writes into outbound stable queues.
There are two basic types of RSI problems:
Those occurring when Replication Server attempts to connect to another Replication Server, and
Those that occur when the locater is invalid.
Although many errors may occur when a primary Replication Server loses its connection to the replicate Replication Server, none of these errors should cause a loss of data or cause the RSI to be suspended. The primary thread continuously attempts to connect to the replicate Replication Server until the connection is reestablished.
Losing the connection terminates the RSI thread at the replicate Replication Server. The thread at a replicate Replication Server exists only while a connection exists. Executing admin who in the replicate Replication Server shows that the primary Replication Server is no longer logged in as an RSI source. When the primary Replication Server reestablishes the connection to the replicate Replication Server, data is replicated across the route to the replicate Replication Server.
An RSI locater identifies the last message that the replicate Replication Server stable queue received from the primary Replication Server. If the RSI locaters at the primary and replicate Replication Servers are not identical, then the primary Replication Server might delete messages from its stable queue at an incorrect location.